avoid costly design mistakes

7/31/2019 Avoid Costly Design Mistakes

1/7

Avoid Costly Design MistakesBy Chip Eskridge, Aker Solutions, Inc., Mike James, DuPont, and Steve Zoller, EnerfabChemicalProcessing.comCommon errors keep plants from getting the most reliable and suitable vessels

Just build it to the Code. Thats the most common response you hear during a design review thatinvolves purchasing a new pressure vessel. Yet, therere as many choices within the ASME codeas there are when searching for your next vehicle. Smart choices can save you money duringfabrication as well as over the lifecycle of the vessel. So, here, well attempt to condense the5,000+ pages (50+ lbs) of the ASME Boiler and Pressure Vessel Code, or the Code, as itsaffectionately known, into a simple guide when specifying vessels, heat exchangers and tanks.Well focus on 10 key factors.

1. Inside diameter versus outside diameter. Process engineers often specify a vessels diameterbased on inside diameter (ID) to ease volumetric calculations. This also will simplifyfabrication/installation of internal hardware (e.g., support rings, trays, distributors, etc.). However,sometimes specifying a vessel based on outside diameter (OD) is better. For instance, after apurchase order is issued, heads are the first things ordered obtaining off-the-shelf heads is more

likely if specified by nominal pipe sizes (NPS), which is OD from diameters 14 in. to 36 in. (For thin-wall heads, i.e., 2-in. thick or less, choosing ID or OD makes little difference, while most headsmore than 36 in. are custom made.) As heads get thicker, hot forming is necessary and dies arebased on ID. Thick, hot formed heads can be OD ordered but require an extra manufacturing step.

In addition, specifying a vessel based on OD has advantages when the project is modeled using 3Dplant design software. If vessel wall thicknesses change during project development, theadjustment is done to the ID keeping nozzle projections, bolt circles, steel design unaffected, etc.

2. Design pressure and temperature. Required wall thickness is more sensitive to pressure thantemperature. Therefore, specifying a design pressure 100 psig over the maximum operatingpressure is more costly than specifying a design temperature 100F higher than whats needed. Adesign pressure of 25 psig50 psig above that at the maximum operating/upset condition and notless than 90% of maximum allowable workingpressure is industry practice. Keep designtemperature to no more than 50F100F abovethat at maximum operating/upset conditions. Also,watch your design pressure and temperature so asnot to cross into the next higher flange class. Check

ASME B16.5 for design temperature and limitationsfor flanges.

3. Vacuum rating. Although current project needsmay not require a vessel to be vacuum rated, overits lifetime, changes in feedstock, products andtechnology will occur. A large number of re-ratescurrently performed are on older vessels originally

not documented for vacuum that now require it dueto process changes. Many new vessels will rate forfull vacuum and all for partial vacuum. So have thefabricator evaluate your proposed design forvacuum and apply it to the code stamp. With todays software, this calculation can be easilyperformed and at no cost. You may get full vacuum rating without any modifications if not,consider spending a little extra now by welding on a stiffening ring and a couple of re-pads to avoidhaving to go through the cumbersome re-rate process and field hydro-test down the road. (Formore on the value of vacuum rating, seewww.ChemicalProcessing.com/voices/plant_insites.html.)

Figure 1 -- Curved heads predominate

and avoid the pressure limitationsof flat heads.Click on illustration for a larger

image.
http://www.chemicalprocessing.com/voices/plant_insites.htmlhttp://www.chemicalprocessing.com/voices/plant_insites.htmlhttp://www.chemicalprocessing.com/Media/MediaManager/Figure1_Design_large.jpghttp://www.chemicalprocessing.com/voices/plant_insites.html


2/7

4. Head choices. Functionality, not cost, should determine head choice; so understanding thefunctional differences is crucial. Dished heads for ASME vessels typically are available in threestyles; elliptical (2:1), flanged and dished (F&D), and hemispherical (hemi-heads). Under 600 psig,elliptical heads are the most common and least expensive in terms of wall thickness and formingcosts. Above 600 psig, hemi-heads are economically attractive due to their inherent low-stressshape; below 600 psig, they are the most expensive choice because they are constructed ofwelded, segmental parts not a single piece. F&D (torispherical) heads have the lowest profile(height/diameter ratio) and compete well with elliptical heads under 100 psig, although they havehalf the volume. The low profile of the F&D head only is advantageous when top head accessibilityis required for maintaining instruments, agitator, etc., or when space is limited below or, forhorizontal vessels, to the sides. For vessels 24 in. or less, off-the-shelf pipe caps (elliptical) provide

the most economical design.

Flat heads have very limited use for pressure vessels more than 24 in. in diameter. Because oftheir flat geometry, they offer far less resistance to pressure than elliptical and F&D heads of thesame thickness. Engineers occasionally will specify a flat head, but this practice is uneconomicalfor pressures above 15 psig25 psig. If a large diameter flat head is necessary for code equipment,

then stiffening the head with structural I-beams is possible but requires sophisticated finiteelemental analysis, a skill that not all fabricators possess.
http://www.chemicalprocessing.com/Media/MediaManager/Table_1_design_large.jpg


3/7

5. Jacket choices. As with heads, consider functionality, not cost. Choosing the correct jacket isparamount to achieve process needs. The three common types conventional, half-pipe anddimple each offer advantages

and disadvantages with respect to process parameters, reliability and cost [1]. Table 1 comparesthem.

6. Cones. Conical sections (cones) are needed where theres a change in diameter or as a bottomhead, e.g., for a bin or hopper. The rule here is keep the transition angle (referred to as the halfapex angle) to 30 or less unless process conditions govern, as exceeding 30 adds costs. The

ASME code demands the piece have a rolled knuckle at both ends when the transition is greaterthan 30; bending stresses complicate the calculation, putting it beyond the skill of manyfabricators.

7. Nozzles loads and projections. The ASME Code [2] requires consideration of all loads. Designersroutinely perform wind and seismic calculations but too often overlook nozzle loads due to thermalpipe stress these can cause visible damage. If attached piping operates at more than 200F wesuggest providing the fabricator with the nozzle loads in Table 2 for a reasonable nozzle stiffening.By providing the fabricator with a reasonable nozzle load, the vessel fabrication and piping designcan proceed in parallel and avoid pipe stress/nozzle loading issues months into fabrication.

Also, nozzle projections below the support ring or lugs shouldnt stick out further than the supportbolt circle or structural steel will have to be removed when setting the vessel. This is ill-advised forheavy equipment.

8. Rectangular tanks. Its not cost effective to specify a rectangular vessel for pressure other thanstatic head; therefore, only consider this configuration for atmospheric tanks. Flat surfaces arehighly stressed under pressure (and vacuum) and the required thickness without adding stiffenerscan be mind-boggling. An engineer needing a rectangular tank often incorrectly specifies API 650or ASME. Neither API 650 nor any other API standard exists for rectangular tanks. Appendix 13 ofthe ASME Pressure Vessel Code provides a methodology but will lead to an expensive over-

design. Most fabricators will apply the stress/strain formulas in Roark [3] to design a safe andeconomical tank that can operate under 15 psig.

9. Cyclic service. If a vessel will experience an unusual number of thermal or pressure cycles overits design life, this could result in premature fatigue failure (usually at a weld) unless preemptivemeasures are taken. Fatigue is cumulative material damage that manifests as a small crack andprogressively worsens (sometimes to failure) as the material is repeatedly cycled. A 1985 surveyshowed that fatigue was the second most prevalentcause of failure in industry (25%), closely behindcorrosion (29%); in the airline industry, itpredominates (61%) [4].

Its up to the purchaser to instruct the fabricator

what design/fabrication practices to follow to avoidfatigue. Cyclic service is usually associated withbatch processes and ASME [5] provides thefollowing rules:

Design for fatigue if N1 + N2 + N3 + N4 400 for non-integral (fillet weld) construction and 1,000 forintegral construction (i.e., no load-bearing fillet welds), or 60 and 350, respectively, in the knuckleregion of formed heads, where N1 is the number of full startup/shutdown cycles; N2 is the number ofcycles where pressure swings 15% (non-integral) or 20% (integral); N3 is the number of thermalcycles with a temperature differential (T) exceeding 50F between two adjacent points no more

Click on illustration for a larger image.
http://www.chemicalprocessing.com/Media/MediaManager/Table_2_design_large.jpg


4/7

than 2.5 (Rt) apart (where R is inside radius of vessel and t is thickness of vessel underconsideration) apply a two-times factor if T exceeds 100F, a four times factor if more than150F, and see Div. 2 for more than 250F; and N4 is the number of thermal cycles for weldsattaching dissimilar materials in which (1-2)T (where is the thermal expansion coefficient)exceeds 0.00034, or for carbon steel welded to stainless steel, the number of cycles where 2Texceeds 340.

Equipment and piping in continuous processes also can experience fatigue due to the relentlessmechanical loading/unloading of reciprocating compressors, piston pumps, bin vibrators or fromvibration, etc., from any type of mis-aligned rotating equipment.

Fatigue failures in welded equipment most commonly occur in fillet welds where theres an abruptchange in equipment geometry. Division 2 of the ASME Code designs around fatigue cracking innozzles by limiting the use of fillet welds. However, fillet welds and sharp corners are ubiquitous inDiv. 1 designs and cant be avoided without cost.

Crack initiation usually begins at the surface due to small microcracks. Therefore, surfacesmoothness is a good defense. Polished surfaces have four times the fatigue resistance [6] butpolishing generally cant be justified for fatigue life alone. Shot peening imparts compressive

stresses into the metal surface that impede crack initiation but, again, only high-end applicationscan economically justify peening. For mid- to small-size process vessels, good weld quality often is themost economical defense against fatigue; so, staterequirements in the equipment specifications.Because fatigue cracks often initiate at the toe orroot of fillet welds, grinding the face to gently blendthe weld into the base metal with a generous radiusremarkably reduces stress risers (Figure 2). Anothermethod to reduce stress risers is to TIG (tungsten-inert-gas) wash a weld toe to improve smoothnessand remove microcracks. Initially target welds wherecyclic loading is occurring. Experience has shown

that most fatigue problems occur due toinadequately supported attachments or wheresaddles/supports lacked wear pads or roundedcorners.

10. Tubing. This can be a significant cost element when ordering large heat exchangers. The costof the tube can vary appreciably depending on the fabrication requirements specified. Its not ourintent to steer you away from the highest quality tube but merely to point out subtleties that cannoticeably affect price.

Diameter. Tubing is specified based on OD. For quickest delivery, stick to commonly stockedsizes, typically -in. and 1-in. tubes for the chemical industry. Specifying smaller tubes (e.g., in.)will increase the exchangers tube count and cost; this will improve duty but will cause higher

pressure drop and may make mechanical cleaning more difficult. Therefore, only consider tubessmaller than in. for cleaner services or when increasing the shell diameter/length isnt an option.Larger tubes (>1 in.) have the opposite effect but may be necessary to satisfy process conditions.

Another option to increase effective surface area without changing tube diameter is to specifyfinned tubes or twisted tubes but those are limited to clean applications.

Figure 2 -- Blend grinding of filletweld can prevent fatigue but the throatof the weld (the distance from faceto root) must meet code requirementsafter grinding.Click on illustration for a largerimage.
http://www.chemicalprocessing.com/Media/MediaManager/Figure_2_design_large.jpg


5/7

Length. Tubes are stocked in 20-ft lengths. Seamless tubes are made from individual billets orhollows and so can vary in length by one to twofeet. The length of welded tubes is more exactbecause theyre produced from a continuousstrip coil. The most wasteful and costly optionfor stocked tubes is ordering units just over 10 ftin length because nearly 50% of the tube isdiscarded. As tube count increases, direct millorders become economically attractive; in suchcases, any length tube can be supplied, if yourschedule allows. Mills have minimum orders(i.e., 2,000 lb.2,500 lb.), though mini-mills willtake orders at half these quantities.

Gauge. Tubes come in different wallthicknesses (or gauge). Industry standards [7]detail the appropriate wall thickness based onmaterial type and service. Table 3 providesguidance for a -in. tube where no prior service

history is available.

Corrosion allowance. This typically isnt added because tubes are considered a replaceablefeature of the exchanger. If designing for a corrosive service, specifying the next-heavier-gaugewall thickness or choosing a higher alloyed tube material is more appropriate.

Seamless versus welded tube. Theres aperception that seamless tubes are more reliablethan welded tubes. This is currently less valid assome manufacturers have developed specializedtechniques for making welded tubing that giveproducts that show no preferential weld corrosionand have properties equal to those of seamless

tubing [8,9]. Seamless tubing will cost more andusually has longer delivery. Welded tubing requiresa greater amount of non-destructive examination(NDE), but this typically only adds pennies per footof tubing if done at the mill [8,9].

Figure 3 -- Welded tubing (left) isinherently more concentric thanseamlesstubing (right). Eccentricity exaggerated

for illustration purposes; A is thegoverning dimension, minimum wallthickness.Click on illustration for a largerimage.


6/7

Eccentricity is inherent in producing seamless tubes [10]. They typically are made by piercing,extrusion or pilgering, generally followed by sizing to obtain final dimensions. The inner mandrel/diecant stay perfectly centered during the tube forming process. Welded tubes on the other handbegin with strip material which is very consistent in wall thickness. So, welded tubes tend to bemore concentric (Figure 3). Seamless tube standards permit larger wall-thickness tolerances thanthose allowed by welded tube standards [11].

Minimum versus average wall thickness. Minimum wall tubes cost a bit more than average walltubing. When its unnecessary to use minimum wall tubing, such as for high pressure or corrosive

service where metal loss is anticipated, it may be more economical to permit the use of averagewall welded tubing and specify additional NDE or corrosion evaluation of the tube seam.

Tube pattern. Shell and tube heat exchangers are the workhorses of the chemical industry. Theseunits typically are fabricated with one of four types of tube patterns 30, 60, 45 and 90 (Figure4). Duty, pressure drop, cleanability, cost and vibration all depend on which pattern is chosen.Consider process needs, not cost, when making the selection.

A 30 or 60 pattern is laid out in a triangle configuration. The main benefit is that approximately10% more tubes can fit in the same area as a 45 or 90 pattern. Theres very little differencebetween the 30 and 60 patterns. Often a thermaldesigner will run analyses of both patterns andselect the one that provides the best pressure drop

and vibration results. The disadvantage of a 30 or60 pattern is that its difficult to mechanically cleanon the shell side. Therefore, such a pattern ischosen for cleaner services; frequently the bundleisnt removable.

The 45 or 90 pattern is selected if shell-sidemechanical cleaning is required. Such a pattern alsorequires a removable bundle. The 45 is morecommon than the 90 because it provides moreshell-side flow disturbance, which improves heattransfer. A 90 pattern is used to reduce pressure drop at the expense of duty and often isemployed in boiling service to enable better vapor disengagement.Make the right choicesIn todays chemical industry, too many engineers given the task of specifying welded equipmentsuch as vessels, heat exchangers and tanks arent well versed in whats necessary to develop aneconomic design that provides suitable safety and performance. Myriad choices must be made and each will incrementally add to the final cost and schedule. When looking for savings, cutting thewrong corners may turn out to be very costly over the equipments service life.

Chip Eskridge, P.E., is principal mechanical/materials engineer for the Aker Plant Services Groupof Aker Solutions, Louisville, Ky. Mike James is a senior consultant, materials engineering, forDuPont, Houston. Steve Zoller is director of fabricated equipment for Enerfab, Cincinnati. Reach

Figure 4 More tubes can fit with 30and 60 configurations but mechanicalcleaning may be harder.Click on illustration for a largerimage.


7/7

them via e-mail [email protected],[email protected]@enerfab.com.

References1. Vessel Design Manual, ITT Industries, White Plains, N.Y. (2000).2. Loadings, Paragraph UG-22, ASME Code, Sect. VIII, Div. 1, ASME, New York City (2007).3. Young, W.C. and R. Budynas, Roarks Formulas for Stress and Strain, 7th ed., McGraw-Hill,

New York City (2002).4. Brooks, C.R. and A. Choudhury, Metallurgical Failure Analysis, McGraw-Hill, New York City(1993).5. Fatigue Screening Criteria, Table 5.9, ASME Code, Sect. VIII, Div. 2, ASME, New York City(2007).6. Deutschman, A.D., W.J. Michels and C.E. Wilson, Machine Design: Theory and Practice,Macmillian, New York City (1975).7. Shell and Tube Heat Exchanger Specification, PIP VESST001, Process Industry Practices,

Austin, Texas (2007).8. James, M.M. and D. ODonnell, When to Specify Welded, Welded and Drawn, or SeamlessTubing, Welding Journal (June 2008). Available online atwww.rathgibson.com/downloads/listing.aspx?did=59.9. James, M.M., Development of High Quality Welded Heat Exchanger Tubing in Lieu ofSeamless, presented at ATI Corrosion Solutions Conference, Sunriver, Ore. (2007).10. Kearns, W.H., Minimizing Wall Thickness Variation in Seamless Tubing, The Fabricator, (Aug.28, 2003), www.TheFabricator.com/TubePipeProduction/TubePipeProduction_Article.cfm?ID=67811. Specifications for General Requirements for Carbon, Ferrite Alloy, and Austenitic Alloy SteelTubes, ASME/ASTM SA/A-450, ASME, New York City (2004).
mailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]://www.rathgibson.com/downloads/listing.aspx?did=59http://www.thefabricator.com/TubePipeProduction/TubePipeProduction_Article.cfm?ID=678mailto:[email protected]:[email protected]:[email protected]://www.rathgibson.com/downloads/listing.aspx?did=59http://www.thefabricator.com/TubePipeProduction/TubePipeProduction_Article.cfm?ID=678

avoid costly design mistakes

Documents